Gastrointest Radiol 17:63-73 (1992)
Radiology ~J Springer-VerlagNew York Inc. 1992
Contribution of CT to Characterization of Focal Nodular Hyperplasia of the Liver Carlo Procacci, ~ Carlo Fugazzola, ~ Marco Cinquino, ~ Gerardo Mangiante, 2 Loretta Zonta, ~ Ivo Andrea Bergamo Andreis, 1 Nicola Nicoli, 2 and Gian Franco Pistolesi 1 D e p a r t m e n t s of ~ Radiology and -' Surgery, University Hospital, Verona, Italy
Abstract. Our personal series of 20 cases of focal nodular hyperplasia (FNH) of the liver is presented. All lesions were studied with computed tomography (CT), 16 of which with surgical control. Retrospective evaluation of the CT features of the identified FNH, along with those of five hepatocellular adenomas (HCA) and 30 hepatocellular carcinomas (HCC), allowed the definition of specific patterns leading to a correct characterization of FNH in 78% of cases. This greatly reduced the diagnostic errors, with the sole exception of patients with fatty liver in whom nuclear medicine may eventually provide a correct characterization. Fine-needle biopsy is thus only necessary in the dubious cases. A precise diagnostic workup of FNH is necessary, since it may avoid the surgical intervention. Key words: Liver, benign l e s i o n s - Liver, focal nodular h y p e r p l a s i a - Focal nodular hyperplasia, CT.
Focal nodular hyperplasia (FNH) is a pseudotumoral lesion of the liver, most often asymptomatic [1-6]. In the past FNH was usually an incidental finding at surgery or autopsy; its preoperative incidence has increased over the last years due to the easier identification by means of the new imaging modalities, notably ultrasonography (US) and computed tomography (CT). The identification of this lesion does not unfortunately mean its proper characterization, since often
Address offprint requests to: Carlo Procacci, M.D,, Department of Radiology, University Hospital, 1-37134 Verona, Italy
problems of differential diagnosis arise vs. either benign or malignant liver tumors [7-10].-A proper characterization is nevertheless important due to the fact that, as for hemangiomas, surgery is not constantly required [3-6, 8, 11, 12]. Moreover, the need to clearly define radiologically this lesion is further justified by the difficulties met by fine-needle aspiration cytology in properly characterizing a FNH [3, 5, 6]. The aim of this paper is to report our personal experience on CT study of 20 FNH identified in 19 patients, in order to assess the reliability of this imaging modality for a proper characterization of the lesion.
Materials and Methods The personal experience refers to 20 FNH identified in 19 patients (one patient presented two lesions; two lesions were associated with an hemangioma): surgical control (either resection or wide wedge biopsy) was performed in 16 patients; the three remaining are being followed, respectively, 30, 45, and 64 months from the initial diagnosis. The retrospective analysis of the FNH identified by CT was done blindly, adding five hepatocellular adenomas (HCA) and 30 hepatocellular carcinomas (HCC). The technique used for their evaluation is the same employed for the assessment of any solid liver mass: the preliminary unenhanced scan is followed by intravenous bolus injection of 30 ml of Conray 60% (Bracco Industria Chimica S.p.A., Milan, Italy) and then rapid drip infusion of 120 ml of the same contrast material. A third control, at the site of the lesion, is performed 5 min after terminating the drip infusion; in some circumstances, late scans (30-40 min) were also obtained.
Results The results refer to the global series of 54 patients presenting respectively FNH (19 of 54), HCA (5 of 54), and HCC (30 of 54).
Fig. 1. FNH: CTfalse negatives. A-D First case. A US (axial scan) shows an ovoid mass (arrows), isoechoic vs. the adjacent normal parenchyma, B On the enhanced CT scan no abnormalities can be recognized within the left lobe. C,D The scintigraphic study shows increased uptake of the radionuclide within the mass (arrows), both on the 99mTc sulfur colloid (C) and on the 99mTcH1DA (D) scans. E-G Second case. The CT study--both the unenhanced (E) and respectively the early (F) and late (G) enhanced scans--shows a large mass within the right liver lobe, with the features of an hemangioma. Only a retrospective evaluation, based on the surgical findings, allows recognition of the FNH (short arrows), with central scar (long arrow), located in front of the portal vein (vp),
Only the 20 F N H are referred to since the other tumors ( H C A and H C C ) (inserted in this series only in order to assess CT characterization capabilities) were constantly identified. CT identified 18 of 20 F N H (90%), with two false-negatives (10%). In one false-negative case, the lesion, 5 c m in diameter and involving the left lobe, had been identified by US (Fig. I A ) a n d was confirmed (Fig. 1C and D) by nuclear medicine (NM). On the CT scan, also at a retrospective evaluation, the lesion could not be identified, being isodense vs. the normal p a r e n c h y m a through all the
C. Procacci et al.: Focal Nodular Hyperplasia
phases of the examination (Fig. 1B). In the other false-negative case, the tumor, located adjacent to a large h e m a n g i o m a , was identified only retrospectively (Fig. 1 E - G ) following the topographic guidelines provided by the surgeon. The m a s s was in fact isodense in all phases of the exam, however, with a central h y p o d e n s e s c a r - - t h e only feature useful for a p r o p e r identification.
We refer to the global series of 54 patients, with exception of the two missed F N H , namely 53 m a s s e s , keeping benign and malignant lesions separated (Table 1). Benign lesions. The diagnosis o f benign lesions was a c h i e v e d in 83% (19 of 23) of the identified benign masses (18 F N H and 5 HCA), notably in 94% of F N H (17 of 18) and 40% of H C A (2 of 5). The proper lesion character&ation was reached in 65% (15 of 23) o f cases (i.e., in 14 of 18 F N H and 1 of 5 H C A ) . As to F N H its characterization could be achieved on the basis of the following features (Table 2).
C. P r o c a c c i et al.: F o c a l N o d u l a r H y p e r p l a s i a
T a b l e 1. A n a l y s i s o f the C T results relating to the characterization o f identified lesions Lesion
No. of cases
B e n i g n lesions (FNH + HCA)
3/23 (13%) --
15/23 (65%) --
T a b l e 2. C T f e a t u r e s in t h e 18 identified F N H , a n d final diagnosis Cases
1 2 3 4 5a 6~ 7 8 9 10 11 12 13 14 15 16 17 18 Total
5 5 4 5 6 3 4 6 4 6 9 5 6 2 7 3.5 5 4
T u m o r density vs. normal parenchyma
< < < < < < = = . < . < < . < = > = (a)
. = = = . . = = .
. < > > > (b)
+ + +
= = .
= = .
= = .
. < > > > (c)
= = =
FNH FNH FNH FNH FNH FNH FNH FNH FNH FNH FNH FNH FNH FNH HCC Benign tumor Benign tumor Benign tumor
< hypodense; = isodense; > hyperdense. (a): < (10); = (7); > (1); (b): < (1); = (14); > (3); (c): < (1); = (14); > (3). " O b s e r v e d in the s a m e patient.
Its densitometric behavior, both on the unenhanced and enhanced scans, was the conclusive sign for diagnosis. On the precontrast scan the mass was hypodense (Figs. 2A, D and G and 3A and E) in 9 of 14 cases and isodense (Fig. 4A) in 5 of 14; in postcontrast scans, both at the early and late controls, the lesion was constantly isodense vs. normal liver parenchyma (Figs. 2B, C, E, F, H, I; 3B, C, F H; and 4B, C). The cen t r a l s c a r was present in 6 of 14 cases, and was hypodense in all on the unenhanced scan (Fig.
2A and D); in two cases it remained hypodense (Fig. 2E and F) and in four hyperdense (Fig. 2C) both at the early and late postcontrast scans (Fig. 2C). The fibrous septa were detected only in 3 of 14 cases. The septa were thick, and were recognized both on the unenhanced (hypodense) and enhanced (hyperdense) scans. Their densitometric features, notably on the enhanced scans, were superimposable to those of scars (Fig. 3A-C). In no cases could persistent hypodensity of the septa be detected. In several lesions small endotumoral structures were
Fig. 2. FNH." CTfeatures. A - C In the left liver lobe the lesion, hypodense with a central scar on the unenhanced scan, turns isodense on the enhanced scan (early and late phase): at the late phase the scar is hyperdense. D - F In the right lobe the mass is hypodense on the unenhanced scan, with a central scar and welldefined margins. The mass turns isodense on the enhanced scan with a hyperdense capsule: the central scar remains hypodense also on the late controls. G - I In the quadrate lobe the lesion, hypodense on the unenhanced scan with thin low-density septa, turns isodense on the enhanced scan; the late control shows a thin capsule and central septa, both hyperdense.
present, with attenuation values similar to those of septa; however, their vascular or fibrous origin was difficult to assess (Figs. 2G-I and 3E-H).
C. Procacci et al.: Focal Nodular Hyperplasia
A hyperdense rim was demonstrated in 5 of 14 lesions on the enhanced scan, notably at the late control (Figs. 2F and 3B, C, G, H). In one case (Fig. 3A-C), FNH, besides presenting gross central fibrous septa, had a thick capsule, appearing hypodense on the unenhanced scan. The indentation on adjacent vascular structures, notably the hepatic veins, played a basic role for the identification of a FNH (isodense vs. the adjacent normal parenchyma in all phases of the CT study) (Fig. 4). The opacification o f the intratumoral biliary ductules, by means of cholangiographic contrast material, was shown in one case (Fig. 3D).
C. Procacci et al.: Focal Nodular Hyperplasia
Fig. 3, FNH: CTfeatures. A - D The right liver lobe is occupied by a large mass which, on the unenhanced scan (A), is hypodense with septa and thick margin. On the enhanced scans (B,C) the lesion turns isodense, with hyperdense septa and capsule. The control, after drip infusion of cholangiographic contrast material (D), shows the presence of hyperdense biliary ductules (arrows) within the lesion. E - H The left lobe is occupied by a large mass, hypodense on the unenhanced scan (E). On the enhanced scans the lesion is isodense both at an early phase (F) and at later controls (5 rain, G; 40 min, H); hyperdense capsule and central septa can be recognized.
Fig. 4. FNH: CTfeatures. The lesion, constantly isodense vs. the adjacent parenchyma, is only identified at a late (5 mini enhanced scan, thanks to the indentation (arrows) on both the intrahepatic vessels and on the inferior vena cava (VC1).
The proper characterization of HCA could be assumed in only one case, which presented with acute abdominal pain and blood loss, due to the identification of hemorrhage within a lesion of the right liver lobe along with a large perihepatic blood collection due to rupture of the tumor (Fig. 6A). The nonspecific diagnosis of a benign lesion was made in 3 of 23 (13%) cases, all FNH (Table 1). Such a diagnosis was made since the tumor was located in a fatty liver (Fig. 5), which made it impossible tc confirm the usual density relation between mass anc normal parenchyma. The wrong diagnosis affected four H c A afid'onc FNH (Table 1). One HCA was diagnosed a FNH since the tumor presented the same attenuatior
Fig. 5. FNH: Nonspecific CT dhtgnoses. A - C Within a fatty liver an evenly hyperdense lesion is recognized, both on the unenhanced (A) and enhanced (B,C) study. Indentation on the hepatic veins, as well as on the inferior vena cava. D,E The mass (arrows) can be recognized only on the enhanced scan (early phase), appearing hyperdense vs. the fatty liver. F A large hyperdense mass on the enhanced scan bulges from the anterior contour of the left lobez diffuse fatty liver.
Fig. 6. HCA: CTfeatures. A A large mass can be recognized within the right lobe, close to the diaphragm, presenting a large central area of hemorrhage and a notable perihepatic hematoma: CT diagnosis of HCA. B,C The tumor is isodense vs. the adjacent
C. Procacci et al.: Focal Nodular Hyperplasia
normal parenchyma both on the unenhanced (B) and enhanced (C) scans; central hyperdense septa: CT diagnosis of FNH.
v a l u e s as a F N H , b o t h o n t h e u n e n h a n c e d a n d e n h a n c e d s c a n s (Fig. 6B a n d C). I n t h e r e m a i n i n g 18% o f c a s e s (4 o f 23) t h e w r o n g d i a g n o s i s w a s c a u s e d b y t h e i n h o m o g e n e o u s p a t t e r n o f t h e l e s i o n , d u e to t h e presence of central hypodense areas, similar to t h o s e m e t in m a l i g n a n t t u m o r s ( F i g . 7): this f e a t u r e w a s d u e , in F N H , t o a c o e x i s t i n g h e m a n g i o m a t o u s a r e a w i t h i n t h e m a s s ( F i g . 7 A a n d B ) ; in H C A it w a s
Procacci et al.: Focal Nodular Hyperplasia
Fig. 7. F N H and HCA: Incorrect CT diagnosis o f malignancy. A,B FNH: The lesion, on the enhanced scan, shows gross hypodense areas surrounded by parenchymal chunks presenting a density superimposable to that of the normal adjacent parenchyma. The following histologic examination (hematoxylin & eosin, • 100) shows the presence of hemangiomatous foci (HHG) surrounded by areas of FNH. C HCA: The tumor is hypodense and inhomogeneous due to a previous hemorrhage. D,E HCA: The tumor, both on the unenhanced and enhanced scans, is characterized by a thick peripheral rim and a large central hypodense area; in this case, the histologic control showed the presence of a large necrotic area.
caused by either a previous hemorrhage (Fig. 7C) or necrosis (Fig. 7D and E).
Malignant tumors. The diagnos& of malignant tumors was achieved in 93% (Table 1) of malignant lesions (28 of 30). The proper tumor character&ation (HCC) was hypothesized in only 30% of cases (9 of 30), all involving patients with liver cirrhosis; the lesions, fairly small in size, were characterized by homogeneous and persistent hypodensity vs. the adjacent parenchyma (the difference in attenuation value, on
the postcontrast scan, was not superior, in these cases, to 20-30 HU). In this group a peripheral hyperdense capsule was constantly identified. The nonspecific diagnosis of malignant tumor was on the contrary achieved in 63% of cases (19 of 30). The diagnosis had to be nonspecific, since the CT features were superimposable to those of other malignant neoplasms (cholangiocarcinoma, metastasis). A wrong diagnosis of FNH was made in 7% of malignant tumors (2 of 30). In one case (fibrolamellar carcinoma) the error was caused by the detection of a central scar, respectively, hypodense on the unenhanced and hyperdense on the enhanced scans; in this tumor the constant, albeit slight, hypodensity vs. the normal liver parenchyma was not adequately assessed (Fig. 8A-C). In the other case, on the contrary, the lesion, slightly hypodense on the unenhanced scan, turned isodense in both the early and late enhanced scans, presenting an uncertain feature of thick capsule; in this mass the postcontrast isodensity of the tumor was due to the lesser attenuation values of the normal liver parenchyma, due to its steatosis (Fig. 8D and E).
C. Procacci et al.: Focal N o d u l a r Hyperplasia
Fig. 8. HCC: Incorrect CT diagnosis o f FNH.
A - C Fibrolamellar carcinoma: The tumor, involving the left liver lobe, is slightly h y p o d e n s e vs. the adjacent p a r e n c h y m a , with a large central scar turning h y p e r d e n s e on the late control. D , E HCC: The t u m o r (arrows), within a fairly fatty liver, is h y p o d e n s e on the une n h a n c e d scan and isodense after contrast materials administration.
could not be corrected even retrospectively; on the contrary, a more thorough evaluation of the fibrolamellar carcinoma would have allowed a correct diagnosis (Fig. 8A-C), whereas in the other circumstance steatosis (Fig. 8D and E) should have led to a nonspecific diagnosis.
The global results, as to FNH, can be summarized as follows. The sensitivity has to be differentiated according to the nonspecific diagnosis of benign lesion or to the specific diagnosis of FNH. In the first circumstance it reaches 85% (17 of 20), being possible also to consider those tumors which (due to the presence of liver steatosis) were not properly characterized; in the second event the sensitivity is lower, namely 70% (14 of 20). The specificity is much higher, namely 91.5% (32 Of 35). On this matter, one has to emphasize that the wrong diagnosis regarding the HCA (Fig. 6B and C)
The preoperative identification of FNH has b e c o m e more frequent over the last years due to the availability of new imaging modalities, notably US and CT [2-5, 7, 11, 12]. A correct characterization of the lesion is useful, since it may avoid surgery; however, it is a goal difficult to reach. Fine-needle aspiration cytology allows differentiation between FNH on the one hand and HCA and well-differentiated HCC on the other hand, based upon the identification of bile duct cells  which are not present in the latter; for the same reason it is not possible to differentiate between FNH and nor-
C. Procacci et al.: Focal Nodular Hyperplasia
mal liver parenchyma [3, 6, 14]. This differential diagnosis can be achieved provided sufficient tissue (core biopsy) is obtained to allow a histologic diagnosis due to the peculiar architecture of FNH. The sonographic pattern of FNH is most often nonspecific, indistinguishable from that of other liver masses, both benign and malignant [7, 12, 15]. The experience gained in the evaluation of FNH by means of magnetic resonance is still too limited to allow a final judgment on its reliability, as well as on the role it may play in the assessment of this pathology [6, 10, 16]. The accuracy of nuclear medicine techniques in the characterization of FNH is high, especially associating 99mTc sulfur colloid and HIDA scintigraphy [10, 12, 15-20]. FNH, characterized by a normal rate of functioning Kupffer cells [2-4], shows an homogeneous, normal or increased, uptake of 99mTc sulfur colloid and 99mTc-HIDA with delayed washout of the latter [17, 18]. By associating the two modalities, FNH can be differentiated from HCC and HCA. In HCC the uptake of 99mTc sulfur colloid or, more rarely, of 99mTc-HIDA is constantly inhomogeneous [19, 21-23]. In HCA, on the contrary, one may have an homogeneous uptake of 99mTc sulfur colloid , the uptake of 99mTcHIDA being inhomogeneous and constantly characterized by fast wash-out . Immediate resort to nuclear medicine after sonographic identification is however not justified, since FNH is a rarer lesion if compared with hemangiomas and malignant tumors, which do require a preliminary analysis by CT, this modality being both able to characterize them [26-28] and to provide, in the case of malignancy, also a judgment of resectability. Only the cases with a dubious CT diagnosis should then be submitted for nuclear medicine along with the rare, already correctly characterized benign lesions (FNH or HCA) for a further diagnostic confirmation. The data in the literature on the reliability of CT in the diagnosis of FNH are limited and often not comparable due to the different modalities of contrast material administration [7, 12, 16]. In our personal experience the CT sensitivity, resorting to the above mentioned technique, reaches 70% (14 of 20 cases) with an even higher specificity (91.5%). By excluding the false-negatives (2 of 20), as well as the FNH associated with liver steatosis (3 of 20), a proper characterization was almost constantly achieved (14 of 15) with one error of overevaluation (diagnosis of HCC). The CT features which in our personal experience contributed to the diagnosis of FNH can be summarized as follows. Attenuation value o f the lesion. The mass appears either hypodense or isodense vs. the normal parenchyma on the unenhanced scan; in all cases it
turns isodense both in the early and late controls after fast contrast material drip infusion. The bolus technique achieves, in the early arterial phase, transient and intense hyperdensity of the mass . This pattern, very useful for the identification of FNH isodense on the unenhanced scan, is nonspecific since HCC also presents the same feature [27, 28]. The isodensity of the lesion on the enhanced scan plays a basic role for a correct diagnosis; similar behavior has in fact been detected in only one case of HCA (Fig. 4B and C). As to the errors of underevaluation among HCC, fibrolamellar carcinoma has constantly been hypodense in the various phases of the study; however, the central scar has been wrongly given a basic diagnostic role. In the second case, the associated liver steatosis (the cause of isodensity of an otherwise hypodense lesion) was not considered. Thus, only steatosis is a severe obstacle to the proper characterization of liver masses, since it cancels the value of the basic sign, namely the comparison between the density of the tumor and the adjacent normal parenchyma both on the unenhanced and enhanced scans [27, 29]. Central scar. This feature, once considered pathognomonic of FNH [3, 4, 6, 19, 30], provides on the contrary a fairly modest contribution to the characterization of the lesion, since it could be identified in only 46% of our cases. Moreover, the same feature can be detected in other tumors, notably fibrolamellar carcinoma [9, 10, 31-35]. The scar is constantly hypodense on the unenhanced scan, being most often hyperdense on the enhanced scan; in some instances it may nevertheless remain hypodense also on the late controls. Capsule and fibrous septa. FNH is most often not properly separated from the normal parenchyma, being without capsule [3, 5, 6, 15, 19, 30]. If present, the capsule is very thin, being recognized only on the enhanced scan as a very tiny hyperdense rim. The fibrous septa are always detected on the precontrast scan as hypodense bands, turning hyperdense on the postcontrast control. Opacification of biliary ductldes. The demonstration of biliary ductules within the fibrous septa by means of cholangiographic contrast material administration with consequent CT control is certainly a basic factor for the diagnosis of FNH [7, 36]; in our experience biliary ducts were opacified in only one of five cases. Indentation on vascular structures. This feature plays an important role whenever the lesion is isodense vs. the normal parenchyma throughout the CT study, devoid of central scar, fibrous septa, or peripheral hyperdense rim.
The indentation on vascular structures can however be demonstrated only if the tumor is more than 3 cm in diameter.
Conclusions Ultrasonography is usually able to identify a FNH. CT is a very useful imaging modality for the characterization of FNH, provided more controls are done during the exam; the only exception are patients with liver steatosis. In the cases with dubious CT features it is mandatory to resort to nuclear medicine, which is not hampered by the diagnostic problems met by CT. The impossibility to properly characterize the lesion makes it necessary to resort to needle biopsy, with the aim of acquiring a histologic diagnosis.
References 1. Edmonson HA. Tumors of the liver and intrahepatic bile ducts. In: Atlas o f tumors. Pathology section. Washington, D.C.: AFIP, 1958:25 2. Ishak KG, Rabin L. Benign tumors of the liver. Med Clin North A m t975;59:995-1013 3. Kerlin P, Davis GL, McGill DB, Weiland LH, Adson MA, Sheedy PF II. Hepatic adenoma and focal nodular hyperplasia: clinical, pathologic and radiologic features. Gastroenterology 1983;84:994-1002 4. Knowles DM, Casarella WJ, Johnson PM, WolffM. The clinical, radiologic and pathologic characterization of benign hepatic neoplasm. Alleged association with oral contraceptives. Medicine 1978;57:223-237 5. Mangiante G, Pistacchi E, Marchiori L, Nicoli N, Dagradi A. Hepatocellular adenoma and focal nodular hyperplasia. Chir Ital 1989;41:117-128 6. Nichols FC, Van Heerden JA, Weiland LH. Benign liver tumors. Surg Clin North A m 1989;69:297-314 7. Mathieu D, Bruneton JN, Drouillard J, Pointreau CC, Vasile N. Hepatic adenoma and focal nodular hyperplasia: dynamic CT study. Radiology 1986;160:53-58 8. Nokes SR, Baker ME, Spritzer CE, Meyers W, Herfkens RJ. Hepatic adenoma: MR appearance mimicking focal nodular hyperplasia. J C A T 1988;12:885-887 9. Rummeny E, Weissleder R, Sironi S, Stark DD, Compton CC, Hahn PF, Saini S, Wittenberg J, Ferrucci JT. Central scar in primary liver tumors: MR features, specificity and pathologic correlation. Radiology 1989;171:323-326 10. Titelbaum DS, Burke DR, Meranze SG, Saul SH. Fibrolamellar hepatocellular carcinoma: pitfalls in nonoperative diagnosis. Radiology 1988;167:25-30 11. Huguet C, Nordlinger B, Baron JC, Parc R, Tubiana JM, Loygue J. Kystes congenitaux et tumeurs benignes du foie de l'adulte. Quelles indications chirurgicales? A p r o p o s de quarante-sept cas. Ann Chir 1986;40:231-235 12. Rogers JV, Mack LA, Freeny PC, Johnson ML, Sones PJ. Hepatic focal nodular hyperplasia: angiography, CT, sonography, and scintigraphy. A JR 1987;137:983-990
C. Procacci et al.: Focal Nodular Hyperplasia 13. Ruschenburg I, Droese M. Fine needle aspiration cytology of focal nodular hyperplasia of the liver. Acta Cytol 1989:33:857-860 14. Klatskin J. Hepatic tumors: possible relationship to use of oral contraceptives. Gastroenterology 1977;73:386-394 15. D'Souza VJ, Sumner TE, Watson NE, Formanek AG. Focal nodular hyperplasia of the liver imaging by differing modalities. Pediatr Radiol 1983;13:77-81 16. Freeny PC. Radiologic diagnosis of hepatic neoplasms. Postgrad Radiol 1990;10:263-283 17. Biersak HJ, Thelen M, Torres JF, Lackner K, Winkler CG. Focal nodular hyperplasia of the liver as established by Tc 99m sulfur colloid and HIDA scintigraphy. Radiology 1980;137:187-190 18. Marabini A, Giorgetti PG, Zamboni M, Cavaggioni M, Braggio P, Nicoli N, Mangiante G. Nuclear medicine identification of hepatic focal nodular hyperplasia: experience of 10 cases. J Nucl M e d Allied Sci 1989;33:154 19. Ros PR, Li KCP. Radiology of malignant and benign liver tumors. Part II: benign liver tumors. Curr Probl Diagn Radiol 1989;18:125-155 20. Welch TJ, Sheedy PF, Johnson CM, Stephens DH, Charboneau JW, Brown ML, May GR, Adson MA, McGill DB. Focal nodular hyperplasia and hepatic adenoma: comparison of angiography, CT, US, and scintigraphy. Radiology 1985;156:593-595 21. Calvet X, Pons F, Bruix J, Bru C, Lomena F, Herranz R, Brugera M, Faus R, Rodes J. Technetium-99m DISIDA hepatobiliary agent in diagnosis of hepatocellular carcinoma: relationship between detectability and tumor differentiation. J Nucl M e d 1988; 29:1916-1920 22. Drane WE, Krasicky GA, Johnson DA. Radionuclide imaging of primary tumors and tumor-like conditions of the liver. Clin Nucl M e d 1987;12:569-582 23. Savitch I, Kew MC, Paterson A, Esser JD, Levin J. Uptake of Tc-99m-Di-isopropyliminodiacetic acid by hepatocellular carcinoma: concise communication. J Nucl Med 1983;24:1119-1122 24. Lubbers PR, Ros PR, Goodman ZD, Ishak KG. Accumulation of technetium-99m sulfur colloid by hepatocellular adenoma: scintigraphic-pathologic correlation. A JR 1987;148:1105-1108 25. Vincent LM, Rho TH, McCartney WH, Mauro MA. Hepatic adenoma. Demonstration of discordant uptake with Tc-99m sulfur colloid and Tc-99m DISIDA. Clin Nucl Med 1984;9:415-416 26. Freeny PC, Marks WM. Hepatic hemangioma: dynamic bolus CT. A JR 1986;147:711-719 27. Freeny PC, Marks WM. Patterns of contrast enhancement of benign and malignant hepatic neoplasms during bolus dynamic and delayed CT. Radiology 1986;160:613-618 28. Itai Y, Ohtomo K, Kokubo T, Yamauchi T, Minami M, Yashiro N, Araki T. CT of hepatic masses: significance of prolonged and delayed enhancement. A JR 1986;146:729-733 29. Lewis E, Bernardino ME, Barnes PA, Parvey HR, Soo CS, Chuang VP. The fatty liver: pitfalls in the CT and angiographic evaluation of metastatic disease. JCAT 1983;7:235241 30. Casarella WJ, Knowles DM, Wolff M, Johnson PM. Focal nodular hyperplasia and liver cell adenoma: radiologic and pathologic differentiation. A JR 1978;131:393-402 31. Bedi DG, Kumar R, Morettin LB, Gourley K. Fibrolamellar carcinoma of the liver: CT, ultrasound and angiography. Europ J Radiol 1988;8:109-112 32. Brandt DJ, Johnson CD, Stephens DH, Weiland LH. Imaging of fibrolamellar hepatocellular carcinoma. A JR 1988;151:295299
C. Procacci et al.: Focal Nodular Hyperplasia 33. Kane RA, Curatolo P, Khettry U. "Scar sign" on computed tomography and sonography in fibrolamellar hepatocellular carcinoma. J Comput Tomogr 1987; 11:27-30 34. Saul SH, Titelbaum DS, Gansler TS, Varello M, Burke DR, Atkinson BF, Rosato EF. The fibrolamellar variant of hepatocellular carcinoma. Its association with focal nodular hyperplasia. Cancer 1987;60:3049-3055 35. Schiebler ML. Hepatocellular carcinoma: MR appearance mimicking focal nodular hyperplasia. A JR 1988;150:472
73 36. Huebener KH, Treught H. Administration of biliary contrast media in computed tomography. In: Felix R, et al., eds. Contrast media in computed tomography. Proceedings of the international workshop, Berlin 1981. Amsterdam: Excerpta Medica, 1981:46-51
Received: April 15, 1991; accepted: May 24, 1991